The application of artificial intelligence (AI) methods, specifically machine learning (ML) techniques and deep learning (DL) neural networks, in manufacturing, can effectively optimize manufacturing processes/workflows by improving decision-making and data analysis, automating complex tasks, and identifying previously unknown patterns. This article discusses AI's benefits, challenges, and key applications in manufacturing.
Image credit: Blue Planet Studio/Shutterstock
Importance of AI in Manufacturing
AI is increasingly becoming crucial to effectively utilize the substantial amounts of data generated by the manufacturing equipment to increase efficiency, accuracy, and safety of manufacturing processes, reduce labor costs, and improve supply-chain efficiencies, which can increase overall productivity, enhance product quality, and reduce downtime.
The adoption of AI has increased among manufacturing firms due to higher revenue volatility, shorter production times, a constant requirement for cost savings, increased inspections and regulations, the need for adaptability and learning on the factory floor, supply chain demands, and manufacturing capacity, and growing need for customized/small-batch goods.
The capability of AI tools to interpret and process the data from the factory floor to detect emerging patterns, identify anomalies in production processes in real-time, and predict and analyze consumer behavior assists manufacturers in gaining comprehensive visibility of all manufacturing operations in every facility across geographies.
Moreover, AI-powered systems can also learn and adapt continuously, significantly improving their performance. For instance, French food manufacturer Danone Group has improved the demand forecast accuracy using ML to reduce forecasting errors, lost sales, and demand planners' workload.
Major Applications of AI in Manufacturing
Quality Inspection: Computer vision can perform several inspection tasks in a restricted environment of a manufacturing facility more efficiently, accurately, and quickly compared to humans.
For instance, a manufacturer of aircraft engines can perform three-dimensional (3D) turbofan blade inspection with micrometer accuracy using computer vision. The system can check several blade properties within a few seconds, allowing the manufacturer to inspect every blade instead of inspecting a random sample.
Moreover, the system can eliminate variations across human inspectors by applying a consistent standard. Automated inspection can also significantly improve consumer product manufacturing efficiency by executing different quality inspection tasks at a substantially faster rate. However, most automated inspection systems are custom-designed for a specific task and cannot be easily retrained for other tasks.
ML-based approaches can be used to develop more flexible inspection systems. Companies such as the BMW Group use automated image recognition for quality inspections and checks and to eliminate pseudo-defects/deviations from the target without defects.
Supply Chain Optimization: AI can monitor and collect supply chain data to identify inefficiencies, predict future demand, and manage inventory effectively. For instance, AI-powered indoor drones can be utilized to monitor warehouse inventories.
Additionally, ML can be employed to predict the demand for manufactured products based on local weather by identifying patterns in the product demand corresponding to the prevailing weather conditions. However, the outcomes of these ML algorithms are typically not easily explainable.
Predictive maintenance and Equipment Monitoring: AI can thoroughly monitor manufacturing equipment through several networked sensors. AI techniques can detect subtle changes, such as slight machine noise and greater than usual vibrations, indicating an impending machine failure from the sensor-generated data.
For instance, industrial product manufacturer Mueller Industries uses an AI-based equipment monitoring system that has successfully detected a problem with bearings on one of the machines, which could have led to significant downtime if not identified and repaired on time.
Thus, AI can potentially facilitate the transition from preventative maintenance to predictive maintenance to avoid machine downtime due to unnecessary preventative maintenance and machine failure, improve worker safety, and lower financial losses.
Advanced Robotics: Although robots have been previously utilized in manufacturing, most were custom-built to perform a specific task repeatedly, cannot be retrained easily, and are blind to their surroundings. AI-powered robots can enable robots to perceive human activities and safely collaborate with humans in a manufacturing facility.
For instance, the application of AI in collaborative robots (cobots) can enable cobots to perceive their surroundings effectively, work more productively compared to humans, improve their adaptability, and reduce overall manufacturing costs.
Japanese automation company Fanuc is utilizing robotic workers to operate its factories continuously. These robotic workers can manufacture crucial components, such as motors, operate all factory floor machinery without breaks, and enable continuous monitoring of all manufacturing operations.
Generative Design: The performance of a product design can be simulated in the real world using AI without physically creating the product. The initial design can be modified based on simulation results to obtain the final/optimal design.
For instance, leading aerospace player Airbus has utilized AI-based generative techniques to manufacture substantially lighter parts than human-designed parts.
Transportation: Autonomous vehicles are deployed extensively in restricted environments such as manufacturing facilities for different tasks, such as moving equipment from one place to another place within the facility, automatically without human intervention.
Leading automotive player Porsche uses autonomous guided vehicles (AGVs) to automate substantial portions of automotive manufacturing. The company is deploying these AGVs to move vehicle body parts from one processing station to the next without involving humans, making their production facilities resilient to sudden disruptions such as pandemics.
Autonomous long-haul trucking is also expected to be crucial in transporting raw materials and finished products to and from manufacturing sites, as navigating the restricted setting of highways is easier than other roadways. Other major applications of AI in manufacturing include order management and process automation. Effective order management requires flexibility in the consumer, market, demand expectations, and manufacturing strategy shifts.
Factories employing AI-based systems for order management can instantly generate purchase requests, efficiently manage the complications of different order types coming from several sales channels, and improve and streamline order management. Similarly, AI-powered software/process mining technologies can enable manufacturers to maintain high production rates indefinitely by detecting and eliminating inefficiencies in the information technology (IT) network.
Challenges of Using AI in Manufacturing
The key challenges involved in implementing AI in manufacturing are the shortage of AI talent, technology interoperability and infrastructure, data quality, real-time decision-making, edge deployments, and trust and transparency. Experienced data scientists and AI professionals are crucial for AI initiatives in the manufacturing sector. However, the shortage of such talent can hamper the efforts to implement AI extensively in manufacturing processes.
Similarly, different production systems, tools, and machines that utilize competing/different technologies result in compatibility issues during AI implementation. Rapid decision-making necessitates real-time prediction services and streaming analytics that allow manufacturers to respond immediately to avoid unintended consequences. Moreover, predictive model deployment on edge devices such as machines, servers, or local gateway is critical for enabling smart manufacturing applications.
Recent Studies
AI is key in low-carbon manufacturing by optimizing industrial structure and production. Thus, understanding the AI-driven low-carbon manufacturing performance is crucial to realizing sustainable resource development and carbon emission reduction targets.
In a paper recently published in the journal Computers and Industrial Engineering, researchers modeled the production process of the manufacturing industry as a network system integrated by the AI application stage, AI technology development stage, and AI upgrade stage for the first time.
Subsequently, an interactive three-stage network data envelopment analysis (DEA) model with ratio data was developed to assess the manufacturing industry in China from 2016 to 2019. Although several regions in China displayed a good performance in the AI application stage, most regions showed a poor performance in the AI technology development and AI upgrade stages.
Additionally, the change trends of the sub-efficiency values in every region differed during the sample period between 2016 and 2019. However, the overall efficiency values of most regions demonstrated a rising trend, which indicated that the AI-driven low-carbon manufacturing industry is rapidly developing in China.
References and Further Reading
Newton, M. (2023). What Unique Benefits Does AI Bring to Cobot Performance? [Online] (Accessed on 25 September 2023)
Crandall, D. J. (2019). Artificial intelligence and manufacturing. Smart Factories: Issues of Information Governance, 10-16. https://policyinstitute.iu.edu/doc/mpi/smart-factories.pdf#page=1
Wuest, T., Weimer, D., Irgens, C., Thoben, K. D. (2016). Machine learning in manufacturing: advantages, challenges, and applications. Production & Manufacturing Research, 4(1), 23-45. https://doi.org/10.1080/21693277.2016.1192517
Liang, S., Yang, J., Ding, T. (2022). Performance evaluation of AI-driven low carbon manufacturing industry in China: An interactive network DEA approach. Computers & Industrial Engineering, 170, 108248. https://doi.org/10.1016/j.cie.2022.108248
Renner, L. A. (2020). How Can Artificial Intelligence Be Applied in Manufacturing? [Online] (Accessed on 25 September 2023)
AI in Manufacturing: Here's Everything You Should Know [Online] Available at https://www.simplilearn.com/growing-role-of-ai-in-manufacturing-industry-article (Accessed on 25 September 2023)
17 Remarkable Use Cases of AI in the Manufacturing Industry [Online] Available at https://www.birlasoft.com/articles/17-use-cases-of-ai-in-manufacturing (Accessed on 25 September 2023)